Product engineering
Laravel products and operational workflows built across interface, backend, and delivery concerns.
LaravelPHPLivewireFilament
Mike Vermeer · Software Engineer
I turn complex workflows into reliable software.
I build reliable products, automation, and AI-enabled engineering workflows—turning ambiguous problems and fast-moving capabilities into dependable systems and better ways of working.
System topology / 01
Software Engineer at Seeders · Nov 2024 — Present.
Internal Laravel and PHP tooling with Livewire and Filament.
Reusable agent skills, repository guidance, code-review workflows, and practical team adoption.
01 / Selected work
Engineering, explained through decisions.
Professional systems, native applications and developer tools—presented through the problems, constraints and trade-offs behind them.
01 / Professional · Confidential clientInternal system
SEO intelligence for a global enterprise client
A custom analytics and automation system that integrated third-party data, processed large datasets, and turned raw metrics into structured insights.
02 / Native macOS productMaintained
EasyHotkeys
A native shortcut utility designed for fast text insertion, thoughtful permission handling, and Keychain-backed privacy.
03 / Native macOS utilityMaintained
NoSleep
A small menu-bar utility built around macOS power assertions, session state, and battery-aware safeguards.
02 / Engineering range
Across the layers that make software useful.
Broad enough to connect the system, deliberate enough to go deep where the problem demands it.
Automation & integration
APIs, data processing, and operational workflows that turn repeated work into dependable, reviewable processes.
API integrationData workflowsAutomationPython
AI-enabled engineering
Reusable agent skills, repository guidance, AI-assisted review, and human-owned validation that turn new capabilities into repeatable practice.
Claude CodeCodexAgent skillsHuman review
03 / AI-enabled engineering
From new capability to dependable practice.
At Seeders, I take a leading role in evaluating new AI capabilities and turning the useful ones into repeatable engineering and operational workflows. The goal is better work, more efficiently—while people retain ownership of decisions, validation, and approval.
Track progress
Follow meaningful advances in models, coding agents, tools, and workflow capabilities while separating durable improvements from novelty.
Assess the opportunity
Define the engineering or business problem, expected value, constraints, and which decisions must remain human-owned.
Encode context
Maintain reusable Claude Code and Codex skills, repository guidance, and engineering standards so agents work with relevant context.
Apply and validate
Use AI across software delivery, code review, research, and selected operational workflows. Validate output through tests, static analysis, runtime checks, and human review.
Standardize and enable
Turn useful lessons into repeatable practices, improve skills and instructions, and help colleagues adopt the workflows effectively.
04 / Approach
Engineering beyond the happy path.
Dependable software comes from clear boundaries, recoverable failures, explicit trade-offs, and systems that remain understandable after launch.
- Clear boundaries
- Recoverable failures
- Explicit trade-offs
- Maintainable systems
Let’s talk
Have an engineering problem worth untangling?
I’m always interested in thoughtful software work and conversations with people who care about how things are built.